Howdy,

I seem to recall an article in Chess Life wherein a Polish computer expert (and expert or master strength Chess player) devised a special algorithim to analyze the King's Gambit.

The results were surprising: The only line where White can draw against "best" play was the 3.Be2 line! I doubt these results mean much for the typical club level player, but wanted to post this. Good luck!

3...f5 is good for Black, since White deprived herself of Qh5.

There's a lot to latch on to here; 
1.e4 e5 2.f4 exf4 3.Be2?! d5 
(3...f5!, as suggested by TN, is better. I'll get back to that.)
4.exd5 Nf6 
(4...c6 may be more precise, e.g. 5.Nc3 Bd6 6. Nf3 Nf6 7.dxc6 Nxc6 8.O-O O-O 9.d4 Re1 or 5...Qh4+ 6.Kf1 Bd6 7.Ne4 Qd8 8.Nxd6 Qxd6 9.Bg4 Ne7 both seem to offer black an appreciable edge.) 
5.Nf3 Bd6
(5...Nxd5 6.O-O h6 (6...g5 7.Re1 c6 8.Nc3 Be6 9.d4 Be7 10.Bd3 Nc7 11.Qe2 Nd7) 7.Nc3 (7.d4 g5 8.Re1 Be7 9.Nc3 c6 10.Bd3 O-O 11.Nxd5 cxd5) 7...Nxc3 8.bxc3 Bd6 9.Bb5+ Nd7 10.Re1+ Kf8 11.c4 may be a little better than the game.)
6.c4 c6 7.d4?!
(Liquidating the doubled pawn by 7.dxc6 Nxc6 seems like a must, and well worth the tempo black gets in exchange.)
7...cxd5 8.c5
(in view of what comes after 9.b4?!, 8.O-O may be a slightly better idea.)
8...Bc7 9.b4? 
(I just don't see the point of this move; trying to create activity on the queenside is too slow. 9.Ne5 g5 10.Nc3 Nc6 11.Bb5 Qe7 looks like a better try, though White is probably still objectively lost.)
9...O-O 10.O-O (10.b5 b6 11.O-O bxc5 12.dxc5 Qe7 13.Bb2 Qxc5 14.Kh1 Nbd7 15.Ba3) 10...a6 11.Ne5 (11.Qb3 Re8 12.Nc3 h6; 11.Nc3 Nc6 12.Qb3 Bf5 13.a4 Be4 14.Rd1 Qe7) 11...g5 12.Nc3 Nc6 13.Nxc6 bxc6 14.a4 Bf5

I'll leave it as an open question: Does any of this look okay for White?

Quote:

   
While 4.e5 is more tenacious than 4.exf5, I'm far from convinced it's enough to overturn the verdict on 3...f5!
After 4.e5 d6 5.d4 dxe5 6.dxe5 Qxd1+ 7.Bxd1 the isolani on e5 appears very weak. 7...Bd7 8.Bxf4 Nc6 9.Nh3 Nge7 10.O-O Nd5 11.Nc3 Nxc3 12.bxc3 and we're nearing an endgame where white has two weaknesses. This should be won for black.
6...Qh4+ 7.Kf1 Bc5 8.Qe1 is also less than satisfactory for White. Consider 8...Qh6 9.Nc3 (or 9.h4 Nc6 10.Nh3 Be3) 9...Be6 10.Nh3 g5 11.Na4 Nd7 12.Nxc5 Nxc5 13.Nf2 Ne4 14.Kg1 O-O-O 15.h4, for example.

TN wrote on 01/09/10 at 06:10:40:


4.exf5? is bad. According to Keres White is doing well after 4.e5 d6 5.d4 dxe5 6.dxe5 Qh4+ 7.Kf1 Bc5 8.Qe1 Qxe1 9.Kxe1 Be3 10.Nc3 or 5...Qh4+ 6.Kf1 Nc6 7.Nf3 Qh6 8.Nc3. I studied these lines 25 years ago and back then I finally agreed with Keres and liked White's position.

Dude, 1981 is so pre-Internet era! There must be plenty of games in databases with this line.